Free
Letters to the Editor  |   July 2014
Response: Observational Study Demonstrates That OMT Is Associated With Reduced Analgesic Prescribing and Fewer Missed Work Days
Author Affiliations
  • Joseph K. Prinsen, DO, PhD
    Department of Medicine, Vanderbilt University Medical Center, Vanderbilt School of Medicine, Nashville, Tennessee
  • Kendi L. Hensel, DO, PhD
    Department of Osteopathic Manipulative Medicine, University of North Texas Health Science Center Texas College of Osteopathic Medicine, Fort Worth
  • Richard J. Snow, DO, MPH
    Ohio Health, Columbus; Department of Family Medicine, Ohio University Heritage College of Osteopathic Medicine at Athens
Article Information
Osteopathic Manipulative Treatment / Pain Management/Palliative Care
Letters to the Editor   |   July 2014
Response: Observational Study Demonstrates That OMT Is Associated With Reduced Analgesic Prescribing and Fewer Missed Work Days
The Journal of the American Osteopathic Association, July 2014, Vol. 114, 530-531. doi:10.7556/jaoa.2014.104
The Journal of the American Osteopathic Association, July 2014, Vol. 114, 530-531. doi:10.7556/jaoa.2014.104
We read with interest the comments of Dr Licciardone1 on our recent publication from the February 2014 edition of the The Journal of the American Osteopathic Association (JAOA) entitled, “OMT [Osteopathic Manipulative Treatment] Associated With Reduced Analgesic Prescribing and Fewer Missed Work Days: An Observational Study.” We are grateful to have the opportunity to respond.2 
We believe that Dr Licciardone's comments fall into 2 broad categories: first, a restatement of the limitations of the article—and of observational studies in general—and second, a summarization of how findings of the OSTEOPAThic Health outcomes In Chronic low back pain (OSTEOPATHIC) Trial relate to our findings. 
We commend Dr Licciardone on his recent publications, including the OSTEOPATHIC Trial; however, we feel obliged to point out that the 2 studies he most frequently cites, and the studies on which he relies most heavily for his comments on our publication, were unpublished3 or in press4 when our manuscript was accepted for publication. Therefore, it would have been impossible for us to reference these studies in our article. We believe that our original publication is corroborated by these studies, with the exception of the pain scores. In response to a critique of an article he authored, Dr Licciardone himself noted in 2013 that research leads to “evolving standards of evidence” that would be difficult to predict if established after the original publication.5 In building an evidence base, each publication is a brick, and together, the bricks establish a solid foundation for clinical decision-making. The authors of each publication reference and build on the best literature available at the time their work is completed. The literature, which is constantly evolving, must periodically be reviewed and summated. 
Dr Licciardone also notes “methodological issues” in our study.1 Observational studies such as ours2 certainly have their limitations, which we believe we discussed thoroughly in our article. Indeed, our limitations section addressed missed data and each potential bias. We encourage readers to review our study2 for a detailed summary of our study's limitations and methodology. 
It is important to note that our study examined the relationship between patients who received OMT by osteopathic physicians and those who did not. Our study did not attempt to compare OMT to allopathic care for low back pain and, therefore, we do not regard Dr Licciardone's comments1 comparing our study to a study with different comparators as relevant or a valid critique. The large number of graduates from colleges of osteopathic medicine in recent years does not logically lead one to conclude that the care residents administer is not representative of the osteopathic medical profession. We agree that the data represent a specific subset of osteopathic physicians, as we discussed originally in our article, and the data should be interpreted accordingly. 
One of the limitations of our study was that only 55% of patients had both initial and final visual analog scores recorded.2 In the corresponding letter to the editor,1 it was noted, “It is unclear why imputation for missing data was not attempted to include the remaining proportion of study patients.” Imputation, or the use of estimated values for missing data, can be attractive to researchers because it is conceptually simple and the resulting sample set has the same number of observations as the complete data set. It can be very appealing when analysis eliminates a large proportion of the data; however, this method of analysis has limitations. Some imputation methods result in biased parameter estimates (eg, means and correlations), unless the data are missing completely at random. The bias is often worse than that with complete-case analysis, especially with the use of imputations based on a mean. The extent to which the bias affects the final analysis is dependent on several factors; yet, all imputation methods underestimate standard errors. It is important to remember that imputed observations are themselves estimates, and their values contain corresponding random error. In light of this fact, imputed values are treated as actual observations for the purpose of statistical analyses. The additional source of error is ignored, resulting in falsely depressed errors and P values. Furthermore, although imputation is possible, it is usually difficult to do well in practice and is not ideal in most instances. Indeed, current Consolidated Standards of Reporting Trials (ie, CONSORT guidelines) discourage the commonly used last-observation-carried-forward method of imputation.6 
Dr Licciardone1 points to the potential bias related to clustering, yet he fails to mention that we addressed this potential bias in our publication.2 Although clustering would theoretically increase the type 1 error rate, it would also, if true, decrease the strength of the statistical difference and imply a false positive finding. Ironically, this theory would suggest that OMT does not alter workdays, prescribing patterns, and pain. 
Although randomized controlled trials (RCTs) may continue to be the gold standard of research, their time and financial commitment places them out of reach for many physicians. Most physicians have limited resources and time, as they are primarily appointed to clinical positions. However, these physicians can meaningfully contribute to medical research by using alternative study designs, such as case studies, observational studies, and retrospective studies. These study designs have an important niche within research, each with their respective strengths and weaknesses. 
Increasingly, funding organizations, such as the Patient-Centered Outcomes Research Institute (http://www.pcori.org/), are focusing on actual patient outcomes rather than results from RCTs. This shift is in response to some of the limitations of RCTs and the generalizability of their results to broad populations. The use of actual patient outcomes in research will increasingly require the use of pseudoexperimental designs to answer questions affecting both policy and payment. Registries such as the American Osteopathic Association Clinical Assessment Program have the ability to contribute in a meaningful way to these decisions, and in some cases registries may be the only way to gain knowledge about health services delivery. In light of this shift, the profession needs to stand behind the accurate, consistent collection of data, which reduces bias introduced by missing data points, and work together to combine results from various studies of osteopathic care in a seamless way for decision makers. This process, which ultimately results in guidelines for care, can ensure that the value of osteopathic care is recognized. 
As scientists, we are obligated to be objective and report findings of our investigations, regardless of whether they corroborate the findings of other studies. Our study demonstrated that patients with low back pain who received OMT had a decrease in the prescription of analgesic medications, fewer nerve-blocking injections, and fewer reported missed or limited-duty days of work. These findings of decreased analgesic medication use in patients who receive OMT concur with previous findings from a major RCT.7 Separating the interaction between use of manipulation, use of pain medications, and patient outcomes of reduced pain and improved functionality would have been difficult with any study design. Our findings of the associations between the use of manipulation and the reduction in pain, as measured by the surrogates of analgesia prescription and increased functionality, reinforce previous findings and demonstrate the value of OMT in managing low back pain. 
References
Licciardone JC. The OSTEOPATHIC Trial demonstrates significant improvement in patients with chronic low back pain as manifested by decreased prescription rescue medication use [letter]. J Am Osteopath Assoc. 2014;114(7):528-529. doi:10.7556/jaoa.2014.103. [CrossRef] [PubMed]
Prinsen JK, Hensel KL, Snow RJ. OMT associated with reduced analgesic prescribing and fewer missed work days in patients with low back pain: an observational study. J Am Osteopath Assoc. 2014;114(2):90-98. doi:10.7556/jaoa.2014.022. [CrossRef] [PubMed]
Licciardone JC, Kearns CM, Minotti DE. Outcomes of osteopathic manual treatment for chronic low back pain according to baseline pain severity: results from the OSTEOPATHIC Trial [published online June 10, 2013]. Man Ther. 2013;18(6):533-540. doi:10.1016/j.math.2013.05.006. [CrossRef] [PubMed]
Licciardone JC, Kearns CM, King HHet al. Somatic dysfunction and use of osteopathic manual treatment techniques during ambulatory medical care visits: A CONCORD-PBRN study. J Am Osteopath Assoc. 2014;114(5):344-354. doi:10.7556/jaoa.2014-072. [CrossRef] [PubMed]
Licciardone JC. Systematic review and meta-analysis conclusions relating to osteopathic manipulative treatment for low back pain remain valid and well accepted. J Bodyw Mov Ther. 2013;17:2-4. [CrossRef] [PubMed]
Moher D, Hopewell S, Schulz KFet al. CONSORT 2010 explanation and elaboration: updated guidelines for reporting parallel group randomized trials. BMJ. 2010;340:c869. doi:10.1136/bmj.c869. [CrossRef] [PubMed]
Andersson G, Lucente T, Davis AM, Kappler RE, Lipton JA, Leurgans S. A comparison of osteopathic spinal manipulation with standard care for patients with low back pain. N Engl J Med. 1999;341(19):1426-1431. [CrossRef] [PubMed]